Data as a Network Effect

Below is the ‘Tip of the Week’ from the Episode 60: Algorithms as a Competitive Advantage, part 2 (Andrew Parker)

We all look for sources of defensibility. How is a startup going to retain its competitive advantage for the long-term? Why can’t someone else just copy an idea? There are a number of sources of defensibility and one frequently cited by Union Square Ventures and mentioned by Andrew Parker in today’s interview, is the network effect. There are even a few different types of network effects. Consider for a second, Facebook. Facebook is not an entirely complicated technology company. Yet, it’s not an idea that can be readily copied. Why is that? Because even if a better version of Facebook launched tomorrow, the users are not going to jump ship for the new network. Their pictures, conversations and most importantly, friend relationships exist with long history, in Facebook. Today we discussed how data can function as a network effect.

Andrew said,

“as a startup gets more data, that value of the startup starts to compound on itself. And the way the data is processed creates more value than the sum of the data itself.”

Recall from the episode with Leo Polovets where he cited the example of Netflix. While Netflix built value based on their key product offering, delivering media in a faster, easier and more informative way to consumers; they can now defend it via the value created from their vast amount of data. So, while Netflix does not have a network effect based on direct friend relationships between users, they do have a indirect user base network effect from the star reviews and ratings.

Imagine a future of ubiquitous data on preferences… not just where a tech company can tell you which movies you are going to like, but one that can tell you:

  • which car to buy,
  • which restaurant to go to,
  • which meal on the menu you’ll like best,
  • which articles to read,
  • which apps to download,
  • what podcasts to listen to…

While the world is in a constant state of change, it’s even conceivable that a smart machine with access to incredible amounts of data will be able to predict what a twenty-year-old’s preferences will be when they are 30, 40 or 50 years old. The examples of advances from big data are limitless. And as long as there are large tech companies developing new algorithms, machine learning and holding the keys to the source data; our network effect dependence on them will only increase.